Ecological Economics 33 2000 353 – 368
SPECIAL SECTION: LAND USE OPTIONS IN DRY TROPICAL WOODLAND ECOSYSTEMS IN
ZIMBABWE A simulation model of miombo woodland dynamics under
different management regimes
J. Gambiza
a,
, W. Bond
b
, P.G.H. Frost
c
, S. Higgins
d
a
Department of Biological Sciences, Uni6ersity of Zimbabwe, Box MP
167
, Harare, Zimbabwe
b
Department of Botany, Uni6ersity of Cape Town, Cape Town, South Africa
c
Institute of En6ironmental Studies, Uni6ersity of Zimbabwe, Harare, Zimbabwe
d
Institute of Plant Conser6ation, Uni6ersity of Cape Town, Cape Town, South Africa
Abstract
Miombo woodlands are crucial to the livelihoods of rural people throughout southern, eastern and central Africa. This paper describes a dynamic simulation model of key ecological processes in miombo and examines the ecological
and economic impacts of various forms of management. The model shows that removing harvestable trees and reducing the level of grazing by livestock causes an increase in grass fuel loads and a corresponding increase in the
frequency of fires. More frequent and intense fires in turn suppress woody regrowth, thereby adversely affecting harvestable tree stocks. Despite the marked ecological response to manipulating the level of grazing, the impacts on
economic performance were minimal. The NPVs for Forestry Commission in particular remained relatively constant under different management regimes. Given these low potential returns, the advantage of applying some of the known
silvicultural management treatments to miombo woodlands seems questionable. Varying the proportion of har- vestable timber trees cut and changing the length of the cutting cycle might suggest that profits to the Forestry
Commission or timber concessionaires could be maximised by harvesting as much timber as possible in a single cutting period. Under such a scenario, however, the woodland would be rapidly converted to bushland. There is a
need to explore further the trade-offs between direct use values, as derived from harvesting and selling timber, and ecological service functions, such as carbon sequestration and modifications of the hydrological cycle. © 2000 Elsevier
Science B.V. All rights reserved.
www.elsevier.comlocateecolecon
1. Introduction
Miombo woodlands, characterised by the over- whelming dominance of trees in the genera
Brachystegia and Julbernardia, cover an estimated 2.7 million km
2
in southern, central and eastern Africa Frost, 1996. They constitute the most
extensive woodland type in Zimbabwe where they provide a wide variety of products and services,
ranging from timber and fuel to medicines and
Corresponding author. 0921-800900 - see front matter © 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 1 - 8 0 0 9 0 0 0 0 1 4 5 - 2
food Campbell et al., 1996. They are also impor- tant for spiritual and cultural reasons, as well as
through their impacts on local climate, soils and hydrological functioning.
Large trees are harvested commercially for tim- ber furniture wood, mine props and charcoal
production, while poles are harvested by commu- nal area residents, mainly for construction and
fuel wood. Information is needed on the extent to which management can maximise the sustainable
productivity of different size classes of preferred species in Zimbabwe mainly Brachystegia and
Julbernardia spp., Pterocarpus angolensis and Afzelia quanzensis. Opening up the tree canopy
increases grass growth Barnes, 1979, thereby increasing the likelihood and intensity of dry-sea-
son fires, a prominent feature of tropical savan- nas. Fire, in turn, suppresses the growth and
recruitment of saplings to the canopy. We there- fore also need to understand more precisely how
tree removal might influence the fire cycle and thereby the processes of establishment and growth
of recruits and the regrowth of harvested trees.
The impacts of fire on woody vegetation are influenced by several interacting factors — timing
and frequency of burning, type of fire, and fire intensity, itself affected by fuel load, particle size
and density, fuel moisture content, and ambient conditions at the time of the fire Barnes, 1965;
Frost et al., 1986; Frost and Robertson, 1987; Chidumayo, 1988; Bond and van Wilgen, 1996.
The impacts vary among woody species and, within species, among size classes. Many species
in miombo are able to resprout from surviving rootstocks, often becoming multi-stemmed in the
process Robertson, 1984. Such plants can be maintained in a multi-stemmed state for decades
by fire and, in some cases, by browsing. We use the term ‘gullivers’ to describe these fire-sup-
pressed multi-stemmed individuals
1
sensu Bond and van Wilgen, 1996. Species that exhibit this
type of recruitment include most of the dominant canopy tree species in miombo such as Brachyste-
gia spiciformis and Julbernardia globiflora, as well as
species such
as Baikiaea
plurijuga and
Colophospermum mopane that are dominant in other regionally important woodland ecosystems.
Fuel load is a key determinant of fire intensity and hence impact on woody plant dynamics.
Standing dead grass and litter make up most of the fuel for savanna fires. Grass production is
positively related to annual rainfall in areas re- ceiving less than 1000 mm Rutherford, 1981; Dye
and Spear, 1982, but is modified by woody plant cover and grazing intensity. The negative relation-
ship between tree biomass and grass production means that fuel loads are generally greatly re-
duced in woodland. This lessens both the proba- bility of the site being able to sustain a fire and
fire intensities, and hence reducing potential dam- age to trees Chidumayo et al., 1996; Frost, 1996.
Likewise, heavy grazing reduces the standing stock of grass, leading in turn to lower fuel loads
and fire intensities.
The key factor limiting recruitment to the woodland canopy is therefore the failure of fire-
suppressed individuals to reach a size at which their bark is thick enough to prevent damage to
the underlying vascular tissues Bond and van Wilgen, 1996. Frequent cutting and, occasion-
ally, browsing by species such as elephant also serve to suppress regrowth. Relatively long dis-
turbance-free intervals are needed to enable indi- viduals to grow to a size that will allow them to
escape into the canopy. Once a sufficient number of trees have reached the canopy, shading reduces
grass production. This in turn leads to less fuel for fires and hence to lower fire frequencies and inten-
sities. Because the relationship between miombo tree canopy cover and grass production is
strongly negatively exponential, there appears to be a threshold that leads either to ‘gulliver’ – grass
mixtures with frequent, high-intensity fires on one side, or to a relatively closed woodland with
fewer, low-intensity fires on the other Frost, 1996. Harvesting canopy trees for commercial or
other purposes may therefore drive a woodland across the threshold towards increased grass pro-
duction, fuel loads, fire frequencies and intensities, resulting in a sharp and difficult-to-reverse change
in vegetation structure and functioning from a woodland to a tall grassland with low-growing,
1
The term ‘gulliver’ alludes to the eponymous character in Jonathan Swift’s novel Gulliver’s Travels who was kept tied
down for a long period by the Lilliputs.
fire-suppressed woody plants. This latter state is sometimes referred to as the ‘fire trap’ Bell,
1984. The main aim of the current work was to
produce a dynamic simulation model of the inter- actions among tree growth, grass production,
grazing, fire and harvesting to assist in under- standing the longer-term dynamics of these wood-
lands under different kinds and intensities of use. The model formed the core of a broader ecologi-
cal – economic model of the people – woodland in- teractions
so prevalent
across much
of south-central Africa Campbell et al., 2000a. In
particular, the model was developed to answer the following questions:
1. What effect does removing harvestable trees have on woodland structure and thereby on
grass fuel loads, and how does this in turn influence fire frequency and intensity?
2. How is woodland structure affected by chang- ing fire frequencies and intensities?
3. What is the sustainable level of commercial timber removal in terms of optimising the
regeneration and regrowth of harvested trees and the recruitment and growth of saplings?
4. What level of grazing by cattle is needed to significantly reduce fuel loads and hence fires?
5. What combinations of livestock and fire are needed to maximise production of commercial
timber? 6. Can management interventions in some of
these ecological processes increase the level of benefit derived from these woodlands by both
commercial and subsistence users?
2. Methods
